Sub - Mains Time Phase Align - Measurement with Time Reference????

[Altsouza]

Hi,

My System is:
- Analog Pre amp main out goes Full Range to Power amp and to Kef R300 Spks. (No way to apply delay to the Mains, only to the Subs)
- Analog Pre amp Sub Out (full range) > MiniDsp > goes to 2 SVS Subs just behind Mains (40cm behind the Mains)
- Use Umik Mic and Outs from a Dac attached to laptop.

I tried to make a Measurement but I got into a problem.
- To measure the Mains is OK. I turn off the Subs, and measure the Mains (Time reference is the Left Spk)
- To measure the Sub I got a problem. I have to disconect the Mains because I want to measure the Subs only , so there is no time reference from L speaker.
Also, is it better to position the Mic at Lp or 1m from each side?

Any help is welcome,

Thanks

Antonio

 

[John Mulcahy]

Best measuring at the listening position. To deal with the reference problem you could use the right as the reference but then shift the result to allow for the difference between left and right. If you measurement position is equidistant from the left and right speakers there should be little or no difference in the timing, however.

 

[Altsouza]

Hi @John Mulcahy
Thanks. They are equidistant so I will try your sugestion.
A newbie question. Each sub is behind the mains and the svs sub already have about 5ms latency, And also goes to miniDsp, so the Subs should have Been some ms delayed relative to the Mains. I can’t delay the mains, only the subs (I use miniDsp for the subs only).
Is this make sense? In this scenario it is possible to align mains and subs? Or am I wasting my time?
Thanks


Antônio

 

[John Mulcahy]

You may find you end up needing to delay the mains relative to the subs to bring things into alignment, if that's not an option you can still use the miniDSP to even out the sub responses. Can also try inverting the subs to see if the alignment is better that way.

 

[Altsouza]

Hi,
I already use Eq and inversion in minidsp, but I wanted to time align Main-Sub, but I see I can't accomplish this without the possibility to delay the Mains. Is it correct?
Thanks
Antonio

 

[jtalden]

You probably are not able to create the conventional timing alignment or the popular reduced group delay alignment. You will probably be able to find a 'delayed alignment' that will still result in good SPL support through the XO range. If there is a SPL dip currently with no SW delay, add 1 ms SW delay at a time and remeasure to see if the SPL support improves. You just need to create enough SPL support in the XO range that it will respond to reasonable EQ. There is no point is going past 5 ms of additional delay. Pick the best SPL support with either SW polarity and keep the additional SW delay to a minimum.

 

[jtalden]

If you like - given an mdat file:
With 3 measurements (acoustic timing is helpful, but not required)

1. L+R
2. SW-L+SW-R
3. All 4 together

Or with the 4 speakers alone (with acoustic timing active)

1. L
2. R
3. SW-L
4. SW-R

I can take a look as see what can be done using the REW alignment tool for the analysis.

 

[Altsouza]

If you like - given an mdat file:
With 3 measurements (acoustic timing is helpful, but not required)

1. L+R
2. SW-L+SW-R
3. All 4 together

Or with the 4 speakers alone (with acoustic timing active)

1. L
2. R
3. SW-L
4. SW-R

I can take a look as see what can be done using the REW alignment tool for the analysis.

Thank you
I will measure.
SW-L+SW-R and L+R - I did not understand this. Can you elaborate?
Thanks
Antônio

 

[jtalden]

You may have misunderstood and I also misunderstood your setup. I was thinking you were high passing you mains. I somehow missed that they are full range in my first read of your post. Since the mains are not high passed (XO is not possible) there may not be any benefit to further delay of the SWs. We may just shift an SPL problem from one bass frequency to another. I am still willing to take a look at your data though to see if there are any benefits to a SW delay increase. Are you now applying a low pass filter to the SWs in the DSP, if so, at what frequency and what filter type are you using?

To try to clarify the measurements needed for the 2 options

For first option:
No timing reference is required - 3 measurements.

1. Measure Both L and R together at the same time (send measurement signal to both L and R with the SWs disabled)
2. Measure both SWs at the same time (send measurement signal to both SWs (with the L and R speakers disabled)
3. Measure all 4 together (REW signal from L+R outputs to stereo receiver inputs thus all 4 speakers are playing the same signal from REW.)

The second option: (this what you may already have measured, if so just post the mdat file)
Acoustic timing must be active - 4 measurements (use John's advice on switching the Reference channel as needed for the measurements).

1. L (alone, with reference on R)
2. R (alone, with reference on L)
3. SW-L (alone, with reference on R)
4 SW-R (alone, with reference on L)

Either option will works okay for my analysis. The order the measurements are taken is not important.

 

[Altsouza]

You may have misunderstood and I also misunderstood your setup. I was thinking you were high passing you mains. I somehow missed that they are full range in my first read of your post. Since the mains are not high passed (XO is not possible) there may not be any benefit to further delay of the SWs. We may just shift an SPL problem from one bass frequency to another. I am still willing to take a look at your data though to see if there are any benefits to a SW delay increase. Are you now applying a low pass filter to the SWs in the DSP, if so, at what frequency and what filter type are you using?

To try to clarify the measurements needed for the 2 options

For first option:
No timing reference is required - 3 measurements.

1. Measure Both L and R together at the same time (send measurement signal to both L and R with the SWs disabled)
2. Measure both SWs at the same time (send measurement signal to both SWs (with the L and R speakers disabled)
3. Measure all 4 together (REW signal from L+R outputs to stereo receiver inputs thus all 4 speakers are playing the same signal from REW.)

The second option: (this what you may already have measured, if so just post the mdat file)
Acoustic timing must be active - 4 measurements (use John's advice on switching the Reference channel as needed for the measurements).

1. L (alone, with reference on R)
2. R (alone, with reference on L)
3. SW-L (alone, with reference on R)
4 SW-R (alone, with reference on L)

Either option will works okay for my analysis. The order the measurements are taken is not important.

Ok, all undestood.
I will make all measurements and post here.
Yes, the mains are full range. I do not wat to mess with them.
I apply a 50hz Low Pass filter to the Subs 24db LR.

Thanks

Antonio

 

[jtalden]

Good, the 50 Hz LR-24 will help the situation.
Either option okay - no need for both.

 

[Altsouza]

Hi @jtalden ,
Here are the measurements. 1st option , no time reference

1. L+R (Mains without Subs)
2. SW-L+SW-R (Subs alone)
3. All 4 together

Correction: The sub Low Pass is 50hz LR-48 (LR-48)
The Sub has minidsp eq turn on , (its turn on on the measurement)
Ths Mains has eq (20 - 150hz range) but is turned off in the measurement, as I use ROON convolver, and I am not using Roon in the measurement.

Thanks

Antonio

 

[jtalden]

I analyzed the file and did not find any increase in SWs delay that improves the SPL using the current SWs polarity. Inverting the SWs polarity and still using the current timing is helpful to smooth the SPL response however. Additional SW delays with the reversed polarity did not help. Below is an overlay chart showing the current setup (green) vs inverting the SWs polarity (red).

I did not find that increasing the delays of the L,R would provide any benefit either so adding that capability to this setup with the current speaker locations would not be helpful. Basically any other timing setups tend to may the SPL worse.

PEQ will help the SPL a lot, but neither of these setups is particularly favorable however for reasonable levels of EQ adjustments. The PEQ would need to apply to both the SWs and L, R to be very effective. Since you only have PEQ for the 2 SWs in the DSP then it is not likely to reduce that 33 Hz peak in any significant way.

You may want to consider changing your speaker and SW locations in the room to find a smoother SPL response as a starting point. That 33 Hz peak is prominent in both the L, R and SWs responses. You may be able to find positions that also help the broad dip at 85 Hz in the L,R.

 

[Altsouza]

I analyzed the file and did not find any increase in SWs delay that improves the SPL using the current SWs polarity. Inverting the SWs polarity and still using the current timing is helpful to smooth the SPL response however. Additional SW delays with the reversed polarity did not help. Below is an overlay chart showing the current setup (green) vs inverting the SWs polarity (red).

View attachment 34489

I did not find that increasing the delays of the L,R would provide any benefit either so adding that capability to this setup with the current speaker locations would not be helpful. Basically any other timing setups tend to may the SPL worse.

PEQ will help the SPL a lot, but neither of these setups is particularly favorable however for reasonable levels of EQ adjustments. The PEQ would need to apply to both the SWs and L, R to be very effective. Since you only have PEQ for the 2 SWs in the DSP then it is not likely to reduce that 33 Hz peak in any significant way.

You may want to consider changing your speaker and SW locations in the room to find a smoother SPL response as a starting point. That 33 Hz peak is prominent in both the L, R and SWs responses. You may be able to find positions that also help the broad dip at 85 Hz in the L,R.

Hi,
Actually I tested inverting the right Sub and apply 2ms delay and got good results.
I will invert the Left sub too.
I have peq in Roon server for the Mains and Sub together. I first apply peq to Subs in miniDsp and after for all using Convolution in Roon.
But this measure I sent you don’t have this Roon eq as I couldn’t measure With it as Rew measurements Did not go into Roon.
Thanks
Antônio

 

[Altsouza]

I am thinking to download a sweep wav file. Play through Roon, and make measurements.
So I will check the results that includes peq I am using.
Does this make sense?
Antônio

 

[jtalden]

EQing the SWs first and then using Roon on both is not likely to be an effective method. At least it is more complicated to find a good result and is likely to result in more PEQ filters than needed. I suggest; first settle on SW and L, R locations, the low pass filter, delay, and polarity that provides the smoothest SPL response, then use Roon PEQ only.

Yes, create the REW measurement file. Use the REW 'From File' measurement option.

Reduce your sweep length to 20-20k Hz for future measurements. There is no value in sweeping below the speakers low frequency capability.

 

[Altsouza]

@jtalden
When saving to file sweep, which option do I choose:
1. Linear / Log or Measurement????
Thanks

 

[jtalden]

Measurement.
See REW: Help/Signal Generator/Measurement Sweep
And
Help/Offline measurements

 

[Altsouza]

Measurement.
See REW: Help/Signal Generator/Measurement Sweep
And
Help/Offline measurements

Thank you